Racial categorization and word identification: The influence of sex, race and regional dialect

Abstract

The speech signal provides information on talker characteristics including socio-ethnic affiliation and racial identity. Regional variation, both similar and divergent from White American English, has been described in African American English. However, it is unknown if such regional dialect variation influences listeners’ racial categorization or word identification accuracy. This work evaluated the influence of listeners’ sex, race, and regional dialect on racial categorization and word identification for Black and White talkers from two dialect regions within North Carolina. Black and White listeners (n = 23) from eastern and central North Carolina participated. In the racial categorization task, listeners heard /hVd/ words produced by male and female Black and White talkers from eastern and western North Carolina. Listeners categorized the perceived talker race for each token as Black or White. In the word identification task, the same listeners matched the speech tokens from the same talkers to one of fourteen /hVd/ words. Results showed an effect of listener sex on word identification accuracy such that female listeners were more accurate than male listeners. No effect of listener race or regional dialect was observed for either task. Follow-up analyses will investigate the interaction between listener and talker sex, race, and regional dialect.